pantherdb / fullgo_paint_update

Update of Panther and PAINT DBs with monthly GO release data
0 stars 0 forks source link

Fix MGI reference ID parsing in update pipeline #57

Open dustine32 opened 2 years ago

dustine32 commented 2 years ago

Not yet sure where exactly this is breaking but the "double-prefix" MGI is causing the whole ID from being output in the go_aggregate view:

annotation_id;accession;term;type;evidence_id;evidence;confidence_code;qualifier;creation_date
1019429424;PTHR21180:AN18;GO:0010875;MGI;1353904743;MGI;IMP;;2020-04-28 10:32:01

Here, both type and evidence are "MGI". For PMIDs, the standard is type="PMID" and evidence="12345".

We need to update queries to properly account for an MGI reference's double-prefix, e.g., "MGI:MGI:12345".